Development of Intention Inference Algorithm Based on EMG Signals at Judging Directional of Arrow Cues

Abstract

Inferring a device wearer’s intention is important for supporting devices that exert force, dexterity and sustainability. In this research, we considered a simple ON and OFF function. If one can discriminate whether an order to move a specific muscle is voluntary or involuntary when the command reaches the periphery, estimating the direction in which the supporting device assists the movement is easy. Therefore, in this study, we measured voluntary biceps brachii muscle contraction in response to a visual stimulus, with various stimulus intensities. We found that the preparation start time and the lifting start time were earlier during the condition (“right directional arrow” and “left directional arrow”) where there was high confidence in the choice to execute a forearm lift, among other available stimuli. In contrast, the center frequency of the biceps brachii muscles, at the time of preparing the lifting motion, tended to be higher. Finally, by using an algorithm to infer movement intention, it was possible to identify if the next movement comprised flexion or extension, before muscle torque generation.